SIGSEGV Error (Due to NULL and non-NULL values TIMESTAMP)

[navistar]

Friends,

While executing parallel jobs to insert/update records on to an Oracle table using Oracle Enterprise stage, I am getting a SIGSEGV error message from the Oracle Enterprise Stage.

Error Message: Operator terminated abnormally: received signal SIGSEGV


This table has 4 date fields, CREATE_DT, MODIFY_DT, EXPR_DT & DELETE_DT apart from other fields. The incoming records have NULL and non-NULL values for each of these date fields and the requirement is to insert the records as is (i.e. date fields having NULL values in input should be populated as NULL and non-NULL values as whatever is coming in).

I have gone through almost all the posts that relate to SIGSEGV error in this forum and tried out various methods mentioned in those posts, but could not find a resolution to my problem. Below I am listing a few of the methods that I have tried:

1. Checked for IsNULL(I/p_Date_Field) and if so, SetNull() else set to input_date_field value
2. Captured NVL() in a stage variable for the input date fields and while writing to the output field, I checked for the value set in the stage variable and if so I SetNull().
3. Split the process into two - first step being the transform step, and writes to a sequential file.(Note: While writing to sequential file, I specified a NULL Field Value.) In the second step, (which is the insert/update step) I used the sequential file created from step 1 and tried to insert/update the records using Oracle Enterprise stage.

None of these worked for me and finally, I dumped Oracle Enterprise stage for Dynamic RDBMS stage to do the same. Eventhough the records were inserted/updated on to the table, using DRS stage came with its own issues. When I used the 'Update Existing Rows or Insert new rows' option in the DRS stage I had to take a big hit on performance for update transactions (which I expected be because default mode is 'Sequential'). So I split the inserts and updates and now the inserts run fast but updates are still a big performance issue.

Can somebody put some thought on this and figure out a way to have this work using the Oracle Enterprise stage? I really hope I can get this issue resolved from the replies to this post, because I am sure a lot of people out there have faced this and need a resolution going forward.

Thanks,
N

[ds_developer]

I had the same problem with parallel jobs in v7.5.2 with NLS enabled. I had to get a patch from IBM to get it to work. The ticket was logged in May 2007 and they provided a patch in early June. Sorry, but I don't have any other details I can give you.

John

[navistar]

Thank you John for that insight.

Would anyone be able to help me know more about this patch from IBM and how to get that?

Cheers!

[ds_developer]

Yes, IBM would be able to give you more information. Call them or your support provider.

John